OpenParable
Skip to content
← Back to skills

Status Card

Design
status-card

Status Card is an Agent Skill for design published in the notwitcheer/hermes-recipes repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

THE way to make a branded image card for a result, finding, leaderboard or quote. Runs a tested local Pillow renderer -> ready PNG. ALWAYS use this for cards; never hand-roll Pillow or HTML (a headless server has no browser).

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/notwitcheer/hermes-recipes/blob/HEAD/recipes/sovereign-gpu/skills/status-card/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/status-card, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Status Card

When to Use

When the operator wants an image card for a result, a finding/stat, a leaderboard or a quote, e.g. to attach to a post. Fully local (Pillow), no cloud, no API key. Cards auto-fit height.

Setup (once)

This skill ships with a dedicated venv pattern: Hermes' own venv is pip-less, so give the skill its own environment with uv:

cd ~/.hermes/skills/creative/status-card
uv venv .venv && uv pip install --python .venv/bin/python pillow

Edit the brand config block at the top of scripts/render_card.py (colors, font path, default footer) once. If the font files are missing the renderer falls back to Pillow's default font instead of crashing.

Procedure

Run the renderer with the terminal tool, using THIS skill's dedicated venv python:

~/.hermes/skills/creative/status-card/.venv/bin/python \
  ~/.hermes/skills/creative/status-card/scripts/render_card.py \
  --type <finding|leaderboard|quote> --title "..." [--subtitle "..."] \
  [--body "..."] [--rows '[["label","value"],...]'] --out /tmp/card.png
  • finding / quote: use --title, optional --subtitle, --body.
  • leaderboard: use --title, optional --subtitle, and --rows (JSON list of [label, value]); row 1 is highlighted. Then attach /tmp/card.png in the reply.

Pitfalls

  • Keep titles short (<~60 chars); the renderer wraps but a very long title looks cramped.
  • Never put the operator's real identity on a card; the footer is the public handle.
  • leaderboard rows are strictly 2 columns ([label, value]) — 3+ columns crash with ValueError: too many values to unpack. For multi-column comparisons use finding with the comparison formatted as plain text lines in --body.

Verification

file /tmp/card.png should report PNG image data, 1200 x <auto-height>.